Trends in cancer incidence and mortality.

Time trends in cancer risk have often been summarised by the observation that mortality from cancers associated with tobacco is increasing rapidly, while mortality from all other cancers is either stable or falling slightly, this slight decline being dominated by the decrease in mortality from stomach cancer. Until recently, and with some variation between the sexes, this simple summary of cancer mortality trends would have been broadly correct for a number of developed countries, and it remains useful in dismissing claims of an impending and unexplained epidemic of cancer, but it does not apply to all countries, and in some there have recently been striking changes in the trends in mortality from cancers associated with tobacco. Cancer mortality has been widely accepted as the most important measure of progress against cancer, since it reflects the impact of cancer on people, and has been considered less subject to distortion than incidence or survival, although this is open to question. Cancer mortality also reflects trends in incidence and survival to a greater or lesser extent. There has been controversy, however, over how cancer mortality trends should be interpreted, as well as over which measures should be used to assess progress in cancer control. An overall summary of trends in mortality from all cancers combined is of limited value in assessing progress against cancer, in any case. Increases in a common lethal cancer may numerically dominate overall mortality trends, perhaps concealing declines in less common or less lethal cancers, while opposite trends in cancers of the lung and stomach, for example, might lead to an overall impression that little has changed. Further, up to a third of cancer patients will not die of cancer, and cancer mortality statistics do not reflect their experience at all. Cancer mortality trends only indirectly reflect trends in the number of people who are diagnosed with cancer in a given year, and those who do die of cancer in a given year may have been diagnosed more than 3 years previously, even though many die earlier: this blurs the responsiveness of routine cancer mortality statistics as a measure of recent progress, and alternative measures have been proposed. Trends in competing risks of death, especially at higher ages, may also complicate the interpretation of cancer mortality trends. The chance of developing cancer, and in that event, the chances of surviving it, are of direct interest to individuals.(ABSTRACT TRUNCATED AT 400 WORDS)